jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steffen Fiedler <steffen.fied...@sourcepark.de>
Subject Re: Remote-Testing Problems
Date Fri, 05 Dec 2003 16:31:24 GMT
Ok, i've got it!
It was a rmiregistry problem with linux:
rmiregistry seems to use /etc/hosts for resolving names
and the machine name was specified for 127.0.0.1
Removing this makes it work.
Thanks anyway,
Steffen


BAZLEY, Sebastian wrote:
>>-----Original Message-----
>>From: Steffen Fiedler [mailto:steffen.fiedler@sourcepark.de]
>>Sent: 05 December 2003 14:13
>>To: JMeter Users List
>>Subject: Re: Remote-Testing Problems
>>
>>
>>BAZLEY, Sebastian wrote:
>>
>>>>-----Original Message-----
>>>>From: Steffen Fiedler [mailto:steffen.fiedler@sourcepark.de]
>>>>Sent: 05 December 2003 12:54
>>>>To: jmeter-user@jakarta.apache.org
>>>>Subject: Remote-Testing Problems
>>>>
>>>>
>>>>Hi,
>>>>
>>>>I want to start JMeter-Tests on several remote computers.
>>>
>>>
>>>Client-server mode is currently not all that efficient.
>>>You will get much better performance running independent 
>>
>>JMeters on each
>>
>>>node in batch mode.
>>
>>I need all testresults in one jtl-file for further processing.
> 
> 
> OK, your choice - I just wanted you to be aware of the performance hit.
> 
> 
>>>>On some machines the test start as expected, on some machines
>>>>nothing happens. The only error-message i detected in JMeter-Log
>>>>(on machine that starts the tests) is:
>>>>
>>>>12/05/2003 1:26:28 PM ERROR - 
>>>>jmeter.engine.ClientJMeterEngine:  java.rmi.ConnectException:
>>>>Connection refused to host: 127.0.0.1; nested exception is:
>>>>	java.net.ConnectException: Connection refused: connect
>>>>java.net.ConnectException: Connection refused: connect
>>>
> [...]
> 
>>>>org.apache.jmeter.engine.ClientJMeterEngine.run(ClientJMeterEn
>>>>gine.java:151)
>>>>	at java.lang.Thread.run(Thread.java:479)
>>>>
>>>>Localhost seems to be wrong but works fine. When i remove the 
>>>
>>>
>>>What do you mean by this?
>>>The log suggests that localhost does not "work fine".
>>
>>If i added localhost to remote-host list, it starts the test,
>>if i add a machine to the list that not starts, i get the message
>>above.
>>For example i have 3 hosts: localhost, machine1 (where the 
>>test starts)
>>and machine2 (where the test doesn't start). The behavior is:
>>
>>remote_hosts=localhost, machine1 -> the test start correctly 
>>on all (2) machines, no exceptions
>>remote_hosts=localhost, machine2 -> test starts only on 
>>localhost, exception above
>>remote_hosts=machine1 , machine2 -> test starts only on 
>>machine1, exception above
>>
>> From 5 machines i tested to remote-start (all debian/java 
>>1.4), only 1 computer
>>behaves as expected and starts to test, others cause the 
>>described exception.
>>I have no idea why, all hosts are listening on 1099 and the 
>>error-message
>>is not comprehensible.
> 
> 
> On the face of it, it looks as though the error is misleading.
> I've had a quick look at the current ClientJMeterEngine code but there is
> nothing obvious to suggest that it might report an error against the wrong
> host.
> 
> What happens if you try:
> 
> remote_hosts=localhost
> remote_hosts=machine1
> remote_hosts=machine2
> 
> separately? [Having started the remote servers as normal]
> 
> Also, there should be some earlier lines in the logfile that might help in
> determining tha cause of the error. [If this is very long, e-mail me
> direct.]
> 
> 
>>>>remote-host which make problems
>>>
>>>>from jmeter.properties, remote-tests runs without exceptions. 
>>>
>>>
>>>As expected.
>>>
>>>
>>>
>>>>All the remote-machines
>>>>are listening on 1099 and there is no output on console and 
>>>>there are no firewalls in use.
>>>
>>>
>>>The error message says that the initial JMeter cannot 
>>
>>connect to the remote
>>
>>>server on 127.0.0.1 = localhost.
>>>
>>>If you are not running a server on localhost, then just 
>>
>>remove it from the
>>
>>>list of remote hosts.
>>>
>>>JMeter will attempt to start all the remote servers that 
>>
>>have been defined.
>>
>>>If I recall correctly, any errors starting remote servers 
>>
>>will cause the
>>
>>>entire run to fail.
>>
>>No, in this case only some servers run the test.
>>
> 
> 
> OK, this may be so. Not sure now if JMeter will try to start further remote
> servers once it has had an error - but this is not really relevant here
> anyway.
> 
> S.
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
> 
> 
> 




---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org


Mime
View raw message